1. The neuromuscular mechanism of sound-production in cicadas has been elucidated by a detailed anatomical and physiological study of Platypleura capitata (Oliv.) and by the analysis of magnetic tape recordings of the song of eight other species in Ceylon.

  2. In all cases the song consists of a succession of pulses, the repetition frequency being between 120 and 600/sec. Each pulse is composed of a damped train of sound waves whose frequency is determined by the natural period of vibration of the tymbals.

  3. A pulse of sound is emitted when the tymbal suddenly buckles or is restored to its resting position by its natural elasticity; in the song of some species both movements are effective. The tymbal muscles, which are responsible for the buckling, have a myogenic rhythm of activity, initiated, but only slightly controlled in frequency, by impulses in the single nerve fibre supplying each muscle. The two tymbals normally act together.

  4. The curvature of the tymbals can be increased by the contraction of accessory muscles, the chief of which are the tensor muscles. This increases the volume of sound emitted at each click and lowers the pulse repetition frequency; the abdomen is raised from the opercula by contraction of the tensor muscles.

  5. The tracheal air sacs form a cavity which is approximately resonant to the frequency of tymbal vibration and can be varied in size by expansion of the abdomen.

  6. Cicada songs, to the human ear, appear to be of great variety. The differences arise largely from the properties of the mammalian cochlea as a frequency analyser; the degree of coherence of phase between pulses, which is probably without significance to the insect, is of great importance in determining the quality of the sound to a human observer. The songs of three species which resemble respectively a bell, a musical phrase and a strident chatter are analysed from high-speed oscillograms, and the difference in quality of sound is explained by reference to the wave-forms.

  7. Some species emit a regular succession of pulses. Others have a slow pattern to their song, produced by the co-ordinated nervous excitation of three functional groups of muscles: (a) wthe tymbal muscles, producing the sound; (b) the tensor muscles, controlling the amplitude and pulse repetition frequency; (c) the muscles controlling the resonance of the air sacs. Of the nine species recorded in Ceylon, those belonging to the genus Platypleura produce their pattern by using (b) and (c), the tymbal muscle being in continuous rhythmic activity; those of the genus Terpnosia use mainly (a) to interrupt the continuity of emission of sound pulses, with some accompanying change in amplitude and pulse frequency. The remaining species use all three muscle groups, but different patterns of co-ordination produce great differences in song.

  8. In one species (Platypleura octoguttata) a distinct courtship song was recorded from a male in close proximity to a female; this ends with attempted copulation.

  9. Preliminary electrophysiological experiments show that the chordotonal sensilla associated with the tympana are extremely sensitive to high-pitched sounds. When the song of another cicada is played back through a loudspeaker the impulse pattern in the auditory nerve corresponds to the pulse modulation envelope, with some after-discharge, as in other insect ‘ears’ (Pumphrey, 1940).

  10. The function of the song is to assemble the local population of a cicada species (males and females) into a small group. It remains to be determined whether it is the main intersexual stimulus in mating behaviour.

According to Myers (1929) the first correct and detailed description of the soundproducing organs of the male cicada was made by Casserius (1600), but credit is usually given to Reaumur (1740), whose account forms the basis for all subsequent accurate descriptions of these organs. Since then a large number of authors (listed by Myers, 1929) have continued the study, and some have attempted to correlate the anatomy with the mechanism of sound-production. There is now general agreement that the spiracular theory of Landois (1867) is incorrect and that, as Reaumur suggested, the sound is produced by the rapid clicking of a pair of drums or tymbals, rib-strengthened chitinous membranes situated on the dorso-lateral surface of what appears to be the first abdominal segment; and that the two tymbal muscles provide the energy for sound-production by pulling on the edges of the tymbals. There are, however, a number of accessory muscles and other structures in the region of the sound-producing organs, and the function of these has been variously interpreted.

Observations on the functioning of the sound-producing organs can be divided into two classes. Felici (1724), Mayer (1877), Lucas (1887) and others experimented by destroying skeletal structures or portions of the musculature and noting the modification of the emitted sound. Myers (1929), Hingston (1922) and others observed living cicadas singing in their natural surroundings and tried to correlate the pattern of the song with movements of the body. Five contributions only in the mass of literature are worth quoting in detail. Distant (1906) provides the standard monograph on the taxonomy ; his names and classification are used throughout this paper. Myers (1928) gives the most complete recent account of the anatomy ; his terminology is largely adopted here. Carlet (1877) gave an account of the anatomy of several European species and attempted to determine the function of the separate muscles and membranes; his detailed results will be discussed where they are relevant to the present work. Lloyd Morgan (1886) made the important observation that the isolated abdomen of Platypleura capensis will ‘sing’ if the tymbal muscle is irritated or stimulated with a ‘weak electric current’, but the significance of this result in a preparation devoid of nervous ganglia seems to have been overlooked both by the observer and by subsequent workers. Finally, Pierce (1948), in the course of an investigation with modem recording apparatus of the songs of various insects, measured the physical characteristics of the song of two American species of cicada.

Species of cicada

The present inquiry was made largely in Ceylon, and is based on an examination of P. capitata (Oliv.), which is common in the parks and gardens of Colombo from March to June. This species and the related P. octoguttata (Fabr.) were used for laboratory experiments. In addition, magnetic tape recordings, field observations and some anatomical investigations were made on seven other species (Text-fig. 1) :

Text-fig. 1.

The nine species of cicada studied in Ceylon. Natural size.

Text-fig. 1.

The nine species of cicada studied in Ceylon. Natural size.

Cicadas do not take readily to captivity and survive best when sleeved on to a growing shrub; in cages they soon damage themselves by their own activity. All crucial experiments were therefore made on specimens captured not more than 48 hr. previously.

Electronic apparatus

Stimulators delivering pulses of controllable frequency, amplitude and duration.

Grass Type P-4 pre-amplifier (condenser-coupled) or d.c. cathode follower (Nastuk & Hodgkin, 1950), with the built-in d.c. amplifiers of the Cossor Type 1049 oscilloscope.

M.S.S. Type PMR/i magnetic tape recorder with Grampion electromagnetic microphone, the high impedance output terminal of the audio-amplifiers being connected direct to the oscilloscope for immediate photographic recording of sounds. These amplifiers have a flat characteristic from 60 to 10,000 cyc./sec. and tests with artificially generated waveforms show that they reproduce pulses at about 5000 cyc./sec. accurately except for the first cycle of the oscillation, which is reduced in amplitude.

Temperature

All experiments were made at the temperature of the laboratory, which was about 30o C.

Anatomy and histology

The plan of the organs of sound-production appears to be the same in all species of cicada of which the males sing. It has been described many times; the best general account is perhaps that in Weber (1930). In describing the Ceylon species it will be convenient to fist the structures in turn and to indicate their relative positions by reference to figures.

Tymbal (Myers, 1928) (Text-figs. 2, 3 and 5)

Text-fig. 2.

Platypleura capitata. Semi-diagrammatic view of muscles and nerves in the region of the organs of sound-production, as seen from within in a specimen bisected longitudinally just to the left of the mid-line. Apart from the thoracic ganglion, gut, blood vessel and the base of the chitinous V, which are shown intact, the remaining structures are duplicated on the left side of the body. The tymbal muscle has been cut across near its base, and the abdominal, auditory and leg nerves are cut near the ganglion. The operculum is not drawn, abd. 1, 2, 3, 1st, 2nd, 3rd abdominal segments; al.p. attachment point of alary muscle; chit. V, base of chitinous V; e.l. elastic ligament; f.m. folded membrane; lat. lateral arm of chitinous V; phr. phragma; III leg, metathoracic leg. Muscles, d.m. dorsal muscles; det.m. detensor tympani muscle ; dors. abd. m. dorsal longitudinal abdominal muscles ; leg.m. basal leg muscles ; lat.abd.m. lateral abdominal muscles; m.int.mes. ‘musculus intersegmentalis mesothoracicus’; spir.m. muscle of the metathoracic spiracle; tens.m. tensor muscle; rent.abd.m. ventral longitudinal abdominal muscles ; vent.m. ventral longitudinal muscles of the metathorax ; w.m. metathoracic direct wing muscles. Nerves, aud.n. auditory nerve; tens.n. tensor nerve; tymb.n. tymbal nerve; (e), nerve to the direct wing muscles; (f) nerve to the anterior dorsal muscles.

Text-fig. 2.

Platypleura capitata. Semi-diagrammatic view of muscles and nerves in the region of the organs of sound-production, as seen from within in a specimen bisected longitudinally just to the left of the mid-line. Apart from the thoracic ganglion, gut, blood vessel and the base of the chitinous V, which are shown intact, the remaining structures are duplicated on the left side of the body. The tymbal muscle has been cut across near its base, and the abdominal, auditory and leg nerves are cut near the ganglion. The operculum is not drawn, abd. 1, 2, 3, 1st, 2nd, 3rd abdominal segments; al.p. attachment point of alary muscle; chit. V, base of chitinous V; e.l. elastic ligament; f.m. folded membrane; lat. lateral arm of chitinous V; phr. phragma; III leg, metathoracic leg. Muscles, d.m. dorsal muscles; det.m. detensor tympani muscle ; dors. abd. m. dorsal longitudinal abdominal muscles ; leg.m. basal leg muscles ; lat.abd.m. lateral abdominal muscles; m.int.mes. ‘musculus intersegmentalis mesothoracicus’; spir.m. muscle of the metathoracic spiracle; tens.m. tensor muscle; rent.abd.m. ventral longitudinal abdominal muscles ; vent.m. ventral longitudinal muscles of the metathorax ; w.m. metathoracic direct wing muscles. Nerves, aud.n. auditory nerve; tens.n. tensor nerve; tymb.n. tymbal nerve; (e), nerve to the direct wing muscles; (f) nerve to the anterior dorsal muscles.

syn. ‘timbale’ (Carlet, 1877), 1 schallplatte’ (Weber, 1930)

Situated on the dorso-lateral surface of what appears to be the first abdominal segment. Detail of the strengthening ribs varies in different species. The apodeme of the tymbal muscle inserts at the dorsal corner, from which a strong forked rib runs down to the middle of the posterior edge of the tymbal and acts as a lever, buckling the other, bowed ribs along a line which runs diagonally across the tymbal through two crescent-shaped markings between the main ribs (Text-fig. 5). The change of shape is therefore complex and there may be several modes of vibration.

Tymbal cover (Myers, 1928) (Text-figs. 1, 3)

syn. ‘tympanal covering’ (Distant, 1906)

An exoskeletal fold projecting forwards on each side of the abdomen and covering the tymbal. This structure is present only in some cicadas and forms the basis of Distant’s (1906) classification, being large in the Cicadinae Dist. (including Platypleura, Rihana, Cryptotympana and Pur ana), small in the Geaninae Dist. (including Terpnosia) and absent in the Tibicininae Dist. (including the famous American seventeen-year locust, Tibicina septemdecim (L.)).

Operculum (Distant, 1906; Myers, 1928; Weber, 1930) (Text-fig. 3)

Text-fig. 3.

Platypleura capitata. Anterior view of a transverse slab including the tymbals and their muscles. The right tymbal cover and the left operculum have been removed, and the dorsal muscles are not drawn.

Text-fig. 3.

Platypleura capitata. Anterior view of a transverse slab including the tymbals and their muscles. The right tymbal cover and the left operculum have been removed, and the dorsal muscles are not drawn.

syn. ‘volet’ (Reaumur, 1740), ‘opercule’ (Cadet, 1877)

An exoskeletal fold of the metathoracic epimeron, projecting backwards on each side ventro-laterally and covering the membranes between thorax and abdomen. The operculum also varies considerably in size in different genera, but its posterior edge is always closely applied to the abdomen, thereby enclosing a space bounded internally by the tympanum and folded membrane. The opercula are rigidly fixed to the thorax and when the abdomen is raised the gap between them and the abdomen is increased.

Chitinous V (Myers, 1928) (Text-figs. 2-4 and 13)

syn. ‘furca’ (Berlese, 1909), ‘entogastre’ (Cadet, 1877), ‘triangle écailleux’ (Reaumur, 1740).

This is a strong skeletal structure forming the basal attachment of the tymbal muscles and projecting dorso-laterally on each side; it is a conspicuous feature of this region of the body and has been variously homologized by different authors. The view taken here, in contrast to all previous workers, is that it is thoracic and is an extreme backwards projection of the metathoracic sternum. This opinion will be elaborated in a later paper (Pringle, 1955) but it is given at once because it affects the view taken of the whole structure of the region.

The arms of the V are firmly attached in the mature adult male to a strong chitinous ridge which forms the posterior edge of the oval in which the tymbal is set, and the arms form a diagonal strut on each side parallel to which runs the tymbal muscle ; the muscle is, in fact, folded round the arms of the V, being kidney-shaped in transverse section in its middle region. The strut is the compression member which balances the tension developed in the tymbal muscle.

Ventrally the chitinous V stands on two short processes which reach and join the exoskeleton of the ventral surface; between these processes run the abdominal nerves (Text-fig. 3). The base of the V is in reality double, internal folds of the skeleton having fused in the mid-ventral line giving a strong basal support for the tymbal muscles.

Folded membrane (Myers, 1928) (Text-figs. 2, 3)

syn. ‘membrane plissée’ (Reaumur, 1740; Carlet, 1877).

A greatly corrugated and opaque membrane separating the air sacs of the thorax from the space enclosed beneath the opercula, and extending on each side from near the midventral line to the lower and anterior edges of the rim of the tymbal. This membrane is stretched when the abdomen is raised.

Tympanum (Myers, 1928) (Text-figs. 2-4)

syn. ‘miroir’ (Reaumur, 1740; Carlet, 1877).

The membrane of the auditory organ (Vogel, 1923)

Air sacs (Text-figs. 3, 4)

The abdomen of the mature male cicada is filled with large air sacs, which compress the viscera and other organs into a narrow dorsal space. The sacs are closely adpressed to the inner surface of the tymbals, folded membranes and tympana, and surround the tymbal muscles except for a narrow connexion along the arms of the chitinous V. The sacs have usually been regarded as tracheal in nature (Snodgrass, 1921), but Myers (1928, 1929), elaborating a suggestion of Hickemell (1923), regards the sac as mesenteric—an expansion of part of the alimentary canal. This is incorrect, at least in Platypleura capitata. The air sacs are paired tracheal enlargements, in communication with the exterior through the metathoracic spiracles.

Text-fig. 4.

Platypleura capitata. Posterior view of the tymbal muscles, tympana and adjacent structures

Text-fig. 4.

Platypleura capitata. Posterior view of the tymbal muscles, tympana and adjacent structures

Text-fig. 5.

Platypleura capitata. External view of the left tymbal, to show details of the strengthening ribs and the line of buckling.

Text-fig. 5.

Platypleura capitata. External view of the left tymbal, to show details of the strengthening ribs and the line of buckling.

Tymbal muscle (Text-figs. 2-4)

The main muscles of the sound-producing apparatus, attached basally to the base of the chitinous V and distally to a thin, approximately circular disk from the centre of which a flat apodeme runs to the tymbal. This muscle is fibrillar in structure.

Tensor muscle (Text-figs. 2, 3)

syn. ‘muscle tenseur de la membrane plissée’ (Carlet, 1877).

A muscle of normal histological appearance running from a well-developed knob on the posterior edge of the metathorax to the anterior rim of the tymbal. Contraction of this muscle has two results. The rim and ribs of the tymbal are further bowed, thereby increasing the tension in the apodeme of the tymbal muscle necessary to click the tymbal, and the whole abdomen is raised, pivoting about its only firm attachment to the thorax in the dorsal region. This movement stretches the folded membrane and increases the space between the ventral surface of the abdomen and the opercula ; the mechanism can be seen from Text-fig. 2. The thoracic knob carries the basal insertion of the tensor muscle so far caudad that it fies behind the anterior rim of the tymbal and the muscle slopes forwards and outwards. The chitinous V is carried with the abdomen in this movement, its connexion with the rest of the metathorax being thin and flexible.

Elastic ligament (Text-figs. 2, 3)

A thin, transparent ligament, stretching from the metathorax to the lower rim of the tymbal, which is stretched by contraction of the tensor muscle.

Ventral muscles (Text-fig. 2)

syn. ‘muscle stemo-entogastrique ‘(Carlet, 1877).

Two muscles on each side running from the anterior part of the metathoracic sternum to the base of the chitinous V. These are antagonists of the tensor and dorsal muscles in their action of raising the abdomen; contraction of the ventral muscles holds the abdomen firmly against the opercula.

Detensor tympani muscle (Text-figs. 2, 4)

A short but very stout band of muscle on each side of the mid-ventral line running between the chitinous V and the sternum of the first abdominal segment (the second abdominal segment of other authors). This is the true ventral intersegmental muscle between thorax and abdomen, but since the chitinous V moves with the abdomen it no longer has the function of depressing the abdomen. Instead it acts as a detensor of the tympanum, bending the strong rim which supports this delicate membrane. In the living male cicada it is always contracted before song commences, thereby creasing the tympanum and presumably protecting it from damage by the high intensity of sound.

Dorsal muscles (Text-fig. 2)

These are four in number on each side; two very short muscles lie within the metathorax and two run from the metathorax to the abdomen. The more dorsal of these probably assist in raising the abdomen. The more ventral is a larger muscle whose posterior insertion is on the rim of the tymbal or on a chitinous disk whence a short apodeme runs to the tymbal rim ; its contraction increases the curvature of the tymbal and it probably acts synergically with the tensor muscle.

Abdominal muscles (Text-fig. 2)

Between each of the segments of the abdomen run, on each side, dorsal and ventral longitudinal muscles which compress the abdomen and lateral muscles which expand it. The latter are attached to a process which projects forward from the next posterior segment.

Flight muscles (Text-fig. 2) (Weber, 1928; Haupt, 1929)

These comprise, on the one hand, the very large indirect muscles of the mesothorax and the muscle called by Weber the ‘musculus intersegmentalis mesothoracicus’, which serves to provide a firm anchorage to the base of the large phragma on the anterior surface of which are inserted the fibres of the mesothoracic longitudinal indirect muscle ; and, on the other, the direct wing muscles, those of the metathorax only being shown in Text-fig. 2. The metathoracic direct wing muscles serve to unfold and fold the hind wings, but appear to be without significance for the wing-beat (see p. 535).

Nervous system (Text-figs. 2, 6)

The thoracic and abdominal ganglia are fused into a single composite ganglion which lies in the mesothorax and gives off a complex bundle of nerves from its posterior end. Among these can be recognized

Text-fig. 6.

Platypleura capitata. Ventral view of the thoracic ganglion showing, on one side, details of the nerves leaving its posterior end. a, abdominal nerves; b, auditory nerve ; c, tensor nerve; d, tymbal nerve; e, nerve to the direct wing muscles; /, nerve to the anterior dorsal muscles; g, metathoracic leg nerve.

Text-fig. 6.

Platypleura capitata. Ventral view of the thoracic ganglion showing, on one side, details of the nerves leaving its posterior end. a, abdominal nerves; b, auditory nerve ; c, tensor nerve; d, tymbal nerve; e, nerve to the direct wing muscles; /, nerve to the anterior dorsal muscles; g, metathoracic leg nerve.

(a) The main abdominal nerves

These nerves run as a single trunk through the basal part of the chitinous V and branch only behind the detensor tympani muscle to supply the second and more posterior segments of the abdomen.

(b) The auditory nerve

This nerve is distinct right from the ganglion but runs with the abdominal nerves through the base of the chitinous V. It diverges in front of the detensor tympani muscles and runs up the posterior edge of the tympanum to the auditory capsule.

(c) The tensor nerve

This leaves the ganglion near the auditory nerve, runs under the ventral muscles to which it sends a branch, and then turns outwards and upwards to supply the tensor and posterior dorsal muscles. Shortly after leaving the ganglion it gives off a fine branch which joins the tymbal nerve; this is probably the fibre supplying the metathoracic spiracle.

(d) The tymbal nerve

This large, single nerve fibre has a separate course from the ganglion, emerging near the auditory nerve and running back a short distance before turning dorsad between the air sacs. Reaching the circular disk at the distal end of the tymbal muscle it separates from the fibre supplying the muscle of the metathoracic spiracle and continues round the tymbal muscle to enter the muscle on its posterior aspect. The single tymbal nerve fibre is 15 /x in diameter ; it branches within the tymbal muscle to supply all its fibres.

(e) Nerve to the direct wing muscles

This emerges from the ganglion more laterally than the nerves already mentioned and runs beneath an internal skeletal process before turning dorsad to the muscles.

(f) Nerve to the anterior dorsal muscles of the metathorax

A small nerve running up the posterior surface of the phragma.

(g) Metathoracic leg nerve

This gives off a branch to the basal leg muscles before entering the leg.

The large fibre of the tymbal nerve can be traced forwards in transverse sections of the ganglion right through its most posterior region to the level at which the fibres of the metathoracic leg nerve break up into neuropile ; here the two tymbal nerve fibres decussate and then break up into small branches. The fibres of the auditory nerve also form a compact bundle which runs forward to the same region. Binet (1894) described briefly the in the ganglion and noted a large pair of cells situated laterally at the level of entry of the leg nerve. These may be the cell bodies of the tymbal nerve fibres.

Analysis of the song

The song of cicadas (Pringle, 1953) consists of a succession of short pulses of sound, each composed of a damped train of sound waves (Text-fig. 7). In P. capitata, in which the song is continuous, the pulse repetition frequency is about 390/sec. and the sound frequency is about 4500/sec. Field records, in which echoes are Hable to affect the waveform of the sound, are not always suitable for high-speed oscillographic analysis, but laboratory experiments with this species show that the amplitude declines approximately exponentially from the start of the pulse (Textfig. 8). The emission of sound is therefore a free oscillation; its frequency is determined by the natural period of vibration of the tymbal. The sound waveform may be sinusoidal {Platypleura) or may show a high proportion of harmonics {Terpnosia stipata, Text-fig. 19); in the latter case the tymbal is evidently vibrating in several modes simultaneously.

Text-fig. 7.

Oscilloscope records of free song. Time marker 50 cyc./sec. A, Platypleura capitata-, B, Terpnosia ransonetti.

Text-fig. 7.

Oscilloscope records of free song. Time marker 50 cyc./sec. A, Platypleura capitata-, B, Terpnosia ransonetti.

Text-fig. 8.

Platypleura capitata. The sound emitted during a single IN-OUT click of the tymbal. This recording was made in the laboratory, and is a reconstruction from four high-speed sweeps of the oscilloscope. Time marker 1 and 5 msec.

Text-fig. 8.

Platypleura capitata. The sound emitted during a single IN-OUT click of the tymbal. This recording was made in the laboratory, and is a reconstruction from four high-speed sweeps of the oscilloscope. Time marker 1 and 5 msec.

In captivity or when handled cicadas emit a series of shrieks, which consist of a short train of pulses, often at high frequency. The work of Myers (1929) suggests that in some species there may be several types of song, but in Ceylon this was observed in only one case, Platypleura octoguttata, in which the normal pattern of song of the isolated male was replaced by a different rhythm and a lower amplitude of sound when it was sitting on a branch next to a female. The full song of the isolated male, which is the sound usually heard in the field, is characteristic of the species and with practice may be used for recognition. There is surprisingly little variation from one individual to another; the whole pattern of neuromuscular coordination is evidently instinctive and inherited with little possibility of modification.

The habits of different species vary. Some, such as Parana campanula, sing only in full sunlight, and the forest quietens at once if a cloud shadow passes. Others, such as Terpnosia stipata, sing in the evenings, when the males descend from the tree-tops to the trunks of the trees. Platypleura capitata sings mainly in sunlight, but a few individuals may be heard late into the evening in Colombo when the population of insects is at its peak.

The rhythm of the tymbal muscle

As was suggested by Reaumur (1740) each pulse of sound corresponds to a click of the tymbal. If in a freshly dead cicada the fine apodeme of the tymbal muscle is pulled steadily with forceps, at a certain critical tension the tymbal clicks from its normal resting position (OUT) to a new position (IN) and emits a pulse of sound. As the tension is slowly reduced a second pulse of sound, sometimes at a slightly different frequency, is emitted as the tymbal clicks OUT again. If the tymbal muscle of P. capitata is electrically stimulated the double nature of the sound pulses is clearly demonstrated (Text-fig. 8). When this species is thus made to produce sound artificially the amplitude of the IN click is always much lower than that at the OUT movement; other species, such as P. octoguttata, have an equal amplitude at IN and OUT. The difference is to be ascribed to the exact form of the buckling which takes place when the tymbal clicks. In the full song of P. capitata the pulses become single, for reasons described later, and it is clear that the pulse repetition frequency is that of the rate of contraction and relaxation of the tymbal muscle.

The mechanism of the rhythmic action of this muscle is the first question to be answered in a physiological analysis of the song. In experiments described in full elsewhere (Pringle, 1954a) it has been established that the rhythm is myogenic; that is, it is determined by the mechanical conditions of loading and elasticity of the tymbal and not by the synchronous arrival of motor nerve impulses. If the tymbal nerve is cut and single stimuli applied to its distal end, the preparation emits a short train of sound pulses, the frequency declining during the train. As many as ten pulses may be emitted in response to a single stimulus in P. octoguttata (Text-fig. 9). Stimulation frequencies of about 120/sec. are sufficient to produce a succession of pulses at a frequency approximating to that of the normal song (P. capitata). The rhythmic mechanism is therefore comparable to that of the indirect flight muscles of Díptera (Pringle, 1949) and Hymenoptera (Roeder, 1951). This similarity in physiological mechanism is one of the arguments which leads to the conclusion (Pringle, 1955) that the tymbal muscle is a modified flight muscle and represents the posterior portion of the vertical indirect wing-muscle complex of the metathorax which has become freed from its connexion with flight through the elaboration of a wing-coupling mechanism and has moved caudad at both ends of its attachment, separating the two functions of flight and sound production which were initially combined in a single mechanism. It was established by experiment that a cicada (P. capitata) in which all the metathoracic and abdominal nerves had been cut could still fly efficiently, proving that the whole of the power for flight is supplied by the large muscles of the mesothorax, which also move the hind wings through their hook coupling to the front pair.

Text-fig. 9.

Platypleura octoguttata. Rhythmic response of the tymbal muscle to a single stimulus applied to the tymbal nerve. Time marker 50 cyc./sec.

Text-fig. 9.

Platypleura octoguttata. Rhythmic response of the tymbal muscle to a single stimulus applied to the tymbal nerve. Time marker 50 cyc./sec.

During full song, in species in which this is continuous, there is evidently a continuous discharge of impulses down the two nerve fibres to the tymbal muscle. Since cicadas will not sing under laboratory conditions this could not be verified directly, but the descending impulse pattern during a short shriek is shown in Text-fig. 10A. This discharge appears as electrical spikes in the tymbal muscle (Text-fig. 10B) and is sufficient to put the muscle into rhythmic activity for a short lime, producing a shriek. The absence from the electromyogram of potential changes at the frequency of contraction of the muscle is discussed by Pringle (1954a); it indicates that the rhythm derives from properties of the contractile myofibrils and not from a rhythmic depolarization of the muscle cell membranes.

Text-fig. 10.

Platypleura capitata. A, spontaneous bursts of motor nerve impulses in the tymbal nerve. Time marker 50 eye./sec. B, spontaneous shriek. 1st trace, electromyogram from the tymbal muscle ; 2nd trace, sound record ; 3rd trace, 50 cyc./sec. ; 4th trace, response of amplifier to 5 mV. square waves. The wave-form of the electromyogram is not accurate, due to overloading and short time constant of the amplifier.

Text-fig. 10.

Platypleura capitata. A, spontaneous bursts of motor nerve impulses in the tymbal nerve. Time marker 50 eye./sec. B, spontaneous shriek. 1st trace, electromyogram from the tymbal muscle ; 2nd trace, sound record ; 3rd trace, 50 cyc./sec. ; 4th trace, response of amplifier to 5 mV. square waves. The wave-form of the electromyogram is not accurate, due to overloading and short time constant of the amplifier.

The small impulses in the record of Text-fig. 10 A are in the motor fibre to the muscle of the metathoracic spiracle, which is maintained tonically closed for long periods.

Synchronization of the two tymbals

Carlet (1877), who had a clear picture of the mechanism of sound production in cicadas, discusses the question of synchronism of action of the two tymbals. He states that in a mature adult cicada the movement of the tymbal during song is too small to be visible, but that in a freshly emerged specimen with incompletely hardened exoskeleton, a considerable distortion takes place. If such a specimen is observed while singing the two tymbals become concave and convex at the same instants. Carlet attempted unsuccessfully to record the tymbal movement on a smoked drum.

Free song records (such as that of Text-fig. 7B) show that there is always a single rhythm of sound pulses ; the two tymbals must therefore be acting synchronously under normal conditions. This implies both that the two tymbal nerve fibres discharge impulses together and, since the rhythm of the tymbal muscle is myogenic, that some further mechanism ensures synchronism of the actual contractions. Text-fig. 11 shows a case in which the synchronism has broken down. The records were made from a fatigued specimen of Rihana mixta held in the hand and stimulated to emit shrieks or single groups of pulses by squeezing the insect; after several minutes of this treatment the two tymbals began to click at different instants and the different amplitudes from the two sides enable the two types of response to be distinguished. They are labelled a and in Text-fig. 11. Sound is emitted at both the IN and OUT clicks of the tymbal, and a third peak is visible on most of the records, due to some peculiarity of tymbal movement in this specimen; each triple pulse group corresponds to one contraction and relaxation of the tymbal muscle. The records are arranged in inverse order ; D was recorded earlier in the experiment than A. In A, B and C each IN-OUT click of the tymbal represents the arrival of one nerve impulse; this is the usual response of a fatigued preparation (Pringle, 1954a). In D, before fatigue is so advanced, the later impulses in a grouped discharge evoke more than one contraction of the tymbal muscle. In A the tymbal nerve fibres on opposite sides are discharging separately; sometimes in B and regularly in C the two tymbals click successively with a constant interval of 6-5 msec. There is evidently a tendency for the two nerve fibres to discharge together or nearly together, suggesting the presence in the ganglion of a macrosynaptic connexion between them which is becoming blocked. This functional connexion normally ensures that impulses occur at the same instant in the two fibres, and only when it is fatigued can they discharge separately. Possibly the synapse is situated where the fibres decussate and run close together, as in the case described by Bullock (1953) in the brain of a polychaete, Protula intestinum; here two giant fibres have a macrosynaptic connexion which also becomes blocked when the preparation is fatigued.

Text-fig. 11.

Sound records from a fatigued specimen of Rihana mixta held in the hand. The pulse groups emitted by the two tymbals are of different amplitude and are labelled α and ² Time marker 50 eye./sec. A, separate contractions of the two tymbal muscles ; B, where a and 3 occur together the interval between them is constant ; C, both tymbal nerves are discharging impulses at the same frequency ; D, fatigue less developed ; repetitive contraction of the tymbal muscles in response to each nerve impulse.

Text-fig. 11.

Sound records from a fatigued specimen of Rihana mixta held in the hand. The pulse groups emitted by the two tymbals are of different amplitude and are labelled α and ² Time marker 50 eye./sec. A, separate contractions of the two tymbal muscles ; B, where a and 3 occur together the interval between them is constant ; C, both tymbal nerves are discharging impulses at the same frequency ; D, fatigue less developed ; repetitive contraction of the tymbal muscles in response to each nerve impulse.

Synchronism of discharge of impulses in the two tymbal nerve fibres still does not explain the synchronization of muscular contractions in the high-frequency myogenic rhythms associated with full song; this must be due to a mechanical coupling between two autorhythmic systems. The explanation of the myogenic rhythm put forward by Pringle (1954 a) is that the sudden release of tension in the muscle when the tymbal clicks IN deactivates the myofibrils and leads directly to relaxation. Probably the fact that both muscles are attached basally to a single skeletal structure produces sufficient mechanical coupling between them to ensure synchronization when both are rhythmically active. A similar explanation was proposed by Pringle (1948) of the synchronism between wings and halteres in Díptera; here again two autorhythmic systems are mechanically coupled through movements of the exoskeleton.

The function of the tensor muscle

Carlet (1877) called this muscle the ‘muscle tenseur de la membrane plissée’, and considered that its function was exclusively to stretch the folded membrane and so increase the volume of sound; he detected a slight diminution in sound intensity when the folded membrane was cut. Undoubtedly the tensor muscle has this function since its contraction raises the abdomen, but it also has a direct influence on the mechanical properties of the tymbal.

Carlet states that there is no tensor muscle of the tymbal and that such is unnecessary since the concavity of this membrane provides the elasticity needed for its recovery when buckled by the tymbal muscle. Experiments were performed on Platypleura capitata to determine whether the tensor muscle increases the curvature of the tymbal and so affects its mechanical properties. The insect was prepared by cutting the nerves just behind the ganglion with a fine pair of scissors inserted through the mesothoracic sternum and removing one operculum to allow electrodes to be placed on one tymbal muscle through the tympanum. Such a dissection upsets the physical properties of the air sacs since one of them is now open to the external air through the tympanum and both through the metathoracic spiracles, whose muscles relax when the nerves are cut. The tymbals and tensor muscles are not, however, disturbed. In the first series of experiments (reported in Pringle, 1954 a) one tymbal muscle was electrically stimulated and pressure was applied with the point of a needle over the distal attachment of the tensor muscle of the same side, simulating its contraction. These showed a definite decrease in the pulse repetition frequency and an increase in the rate of relaxation of the tymbal muscle after each IN click : results which would be expected if the curvature and therefore the elastic restoring force of the tymbal were increased. In the second series of experiments high frequency (400/sec.) electrical stimuli were applied through a second pair of electrodes to the tensor muscle while the tymbal muscle was strongly excited. The results are shown in Text-fig. 12. At the start of stimulation of the tymbal muscle the pulse repetition frequency was 278/sec. (A); when the tensor muscle contracted the frequency dropped to 239/sec. (B). As stimulation of the tymbal muscle was contined the frequency rose steadily until it reached 440/sec. (Q ; tensor contraction now reduced this to 408/sec. (D). Again it is clear that the tensor muscle, by increasing the curvature of the tymbal and raising the critical tension which must be re-developed in the tymbal muscle before the tymbal clicks IN, has lowered the frequency of the myogenic rhythm.

Text-fig. 12.

Platypleura capitata. Sound records from a laboratory experiment illustrating the effect of contraction of the tensor muscle. Time marker 1 and 10 msec. A and C, tensor muscle relaxed. B and D, tensor muscle contracted. For further explanation see text.

Text-fig. 12.

Platypleura capitata. Sound records from a laboratory experiment illustrating the effect of contraction of the tensor muscle. Time marker 1 and 10 msec. A and C, tensor muscle relaxed. B and D, tensor muscle contracted. For further explanation see text.

A further effect is noticeable. When the tymbal muscle of P. capitata is stimulated the sound pulses are always double, both the IN and OUT clicks of the tymbal emitting sound, with the IN click smaller in amplitude than the OUT (Text-figs. 8, 12-4). When the tensor muscle contracts, the amplitude of sound at the IN click increases until it is as large as or larger than that at our (Text-fig. 12B). The tension required to click the tymbal IN is greater and, when it does click, more energy is dissipated. There is no measurable change in the sound frequency, but the peak amplitude increases about 15%.

These experiments show that the tensor muscle does control the mechanical properties of the tymbal and that its contraction, while the tymbal muscle is rhythmically active, increases the volume and duration of the emitted sound. The volume which could be produced in the laboratory with this preparation was never comparable to that of the full song in the field and other factors must also be involved. One of these is no doubt the resonance of the air sacs (see later), but it may be that contraction of the lower dorsal muscle, which has a tensor action, is also needed to achieve the maximum sound output. This muscle is awkwardly placed for stimulation in an intact insect.

Another function of the tensor muscle, which results from its action in increasing the elastic restoring force of the tymbal, is to balance the tonic tension which is liable to build up in the tymbal muscle when it has been rhythmically active for some time. In the absence of simultaneous tensor contraction this tension is ultimately sufficient to cause a failure of the sound-production mechanism, since, after a time varying from a few seconds to several minutes depending on the state of the preparation, the tymbal may suddenly fail to click OUT again after an IN click; when this occurs the tymbal muscle goes into a tetanus and the rhythmic cycle cannot re-start until stimulation has ceased and re-commenced. The phenomenon was never observed when the tensor muscle was also excited. The physiological implications of the IN failure are discussed by Pringle (1954 a).

It seems very probable that the tensor muscles are tonically contracted in nearly all species of cicada when they are singing in the field. In some species (see later) analysis of the pattern of song suggests that it is produced by periodic contraction and relaxation of this muscle but, in Platypleura capitata and Terpnosia ransonetti (Pl. ii, Text-fig. 7, Table 2), where there is no pattern, the apparent singleness of the sound pulses indicates that they are primarily emitted at the IN click of the tymbal, and that the smaller OUT click is lost in the powerful vibration resulting from the IN click.

The nature and function of the air sacs

The tracheal nature of the large abdominal air sacs, which was denied by Myers (1928, 1929), has been clearly established. Dissection of fresh specimens showed that they are lined by a membrane which is not wettable by water on the inner (air) surface, but are readily filled with melted paraffin wax when a freshly dead cicada is immersed in this liquid in a flask from which the air has been removed by a suction pump. Experiments with xylol coloured with Sudan III showed that the cavity of the air sac is in communication with the external air through the metathoracic spiracle and that the tymbal muscle derives its tracheal supply from the same spiracle. Injection of coloured xylol into the first abdominal spiracle (lying in the ring of that segment bearing the tympanum (Text-fig. 4)) reddens a tracheal trunk running dorsad to the viscera and sending a branch under the tymbal muscle, but does not colour the fibres of the muscle itself.

The air sacs must therefore be regarded as of metathoracic origin, their position in the abdomen resulting from their great posterior enlargement. Some light was shed on the manner of their formation by examination of a freshly emerged male cicada in which the walls of the sacs were still folded and partially collapsed. The abdomen was largely occupied by a liquid-filled body cavity, and stretching from the air sacs were numerous very small striated muscle fibres running to the body wall. When they have just emerged, cicadas excrete a large quantity of fluid, which is ejected from the rectum in the form of a fine rain. It is probable that the enlargement of the air sacs takes place partly through the contraction and subsequent absorption of the small muscle fibres, and partly by the excretion of body fluid which draws the walls of the sacs closely into apposition with the body wall, the tymbal muscle and the other internal organs.

Occasionally specimens were found in which there appeared to be a dorsal median air sac in addition to the two lateral ones. This median cavity has a wettable wall and seems to be the true body cavity into which air has been admitted. Only two out of about 100 specimens examined (Platypleura capitata) showed this cavity, and it is possible that its formation was due to an accidental rupture of the delicate lining of the tracheal sacs. Both insects were capable of song.

Paraffin filling of the normal lateral air sacs was used to make a wax cast in order to determine their size and shape (Text-fig. 13). Physically they can be considered as a single cavity, since the membrane dividing them is extremely delicate in the mature insect. This cavity has, on each side at the front end, two openings; a dorsal opening through the tymbal, covered in Cicadinae by the tymbal cover, and a ventral opening through the tympanum covered by the operculum. Experiments were carried out to determine its resonant frequency in a freshly dead cicada, and measurements were also made on a Perspex cavity moulded round the wax cast. The method found most successful was to excite the cavity with a pulsed air jet applied across one tymbal. The sound emitted was picked up with the microphone and observed on the oscilloscope whose time-base was arranged to be triggered by the sound-wave of greatest amplitude. Successive pulses of sound were thus superimposed on the trace and the effect of extraneous noises reduced to a minimum. Specimen records are shown in Text-fig. 14. The frequency of sound emitted in full song by P. capitata varies in different individuals from 4400 to 4700/sec. Measurements of the resonant frequency of the Perspex cavity moulded round a wax cast of the air sacs gave a value of 4220/sec (± 57 ; four determinations). The values for a freshly dead cicada under various conditions of mutilation are given in Table 1.

Text-fig. 13.

Platypleura capitata. Outline drawings of a wax cast of the cavity of the air sacs.

Text-fig. 13.

Platypleura capitata. Outline drawings of a wax cast of the cavity of the air sacs.

Text-fig. 14.

Platypleura capitata. Determination of the resonant frequency of the air sacs. Superimposed oscillograms of the sound emitted when the cavity is excited with a pulsed air jet. A, intact insect; B, tymbal covers removed; C, tymbals removed; D, opercula removed.

Text-fig. 14.

Platypleura capitata. Determination of the resonant frequency of the air sacs. Superimposed oscillograms of the sound emitted when the cavity is excited with a pulsed air jet. A, intact insect; B, tymbal covers removed; C, tymbals removed; D, opercula removed.

TABLE I.

Resonant frequencies of the air sacs of a dead cicada (Platypleura capitata)

Resonant frequencies of the air sacs of a dead cicada (Platypleura capitata)
Resonant frequencies of the air sacs of a dead cicada (Platypleura capitata)

The method achieves no great accuracy, and the differences between the measurements under the first three conditions are not significant at the 5 % level. Those between the two positions of the abdomen are just significant and the remaining differences clearly so. The biggest effect is produced by removing the opercula, which normally form the lower boundary of the cavity. All the values are higher than those obtained from the Perspex cast of the cavity. Probably the contact between the opercula and the abdomen in a dead specimen is less good than in the living insect (in which the ventral muscles may be tonically contracted) and the resonant frequency is thereby raised ; in the cast the cavity is completely sealed at this gap and the frequency is too low. One may conclude that adjustment of the gap in the living insect by means of the accessory musculature provides a means of tuning the cavity to the natural frequency of vibration of the tymbal and so of increasing the volume of the emitted sound. Dilatation of the sacs by expansion of the abdomen appears to have a smaller effect, but might raise the ‘Q ‘ (sharpness of tuning) of the cavity and so reduce the damping.

The function of the tymbal covers is less easy to understand. They may be purely protective. There is no obvious difference in the efficiency of song between cicadas of the genera Platypleura, in which they are large, and Terpnosia, in which they are small. There is a difference (which may not be general as between Cicadinae and Geaninae) in the quality of the sound and in the manner in which the pattern of the song is generated, but it is not at all obvious how this can be correlated with the presence or absence of tymbal covers.

It is worth mentioning here that air sacs are absent in Fulgoromorph and Jassidomorph Auchenorhyncha, where Ossiannilsson (1949) has shown that the tymbals and tymbal muscles are present, sometimes in both sexes. The song of these small insects consists of a succession of pulses, which are evidently produced, as in cicadas, by the clicking of the tymbals. The damping of the tymbal vibration is, however, high, and no appreciable volume of sound is generated.

Summary of the mechanism of sound-production

  1. The song of cicadas consists of a succession of damped pulses of sound produced by the free vibration of the tymbals as they click either IN or OUT between their resting and a buckled position. The rhythm of tymbal movement is maintained by a myogenic rhythm of contraction and relaxation of the tymbal muscles. The occurrence (but only to a small extent the frequency) of this rhythmic activity is controlled by a single nerve fibre supplying each tymbal muscle; impulses in the two nerve fibres normally occur together, and there is evidence of a macrosynaptic connexion between them in the ganglion.

  2. Of the associated accessory muscles, the tensor muscles (and possibly also a dorsal muscle) by their contraction increase the curvature of the tymbals and so increase the sound output, especially at the IN click. There is an accompanying reduction in the repetition frequency of the sound pulses and a raising of the abdomen in relation to the opercula borne on the thorax.

  3. The air sacs, which are tracheal in nature and open to the exterior through the metathoracic spiracles, form a cavity which is at least approximately resonant to the frequency of tymbal vibration. The resonant frequency is altered by expansion of the abdomen (by contraction of the lateral abdominal muscles) and by raising it, which increases the gap between the abdomen and opercula.

The quality of the song to the human ear

Pumphrey (1940) has pointed out that the quality of many insect songs to the human ear may bear little relationship to its significance to another insect. The mammalian cochlea is a frequency-analysing device, each auditory nerve-ending responding maximally to a particular audio-frequency (Galambos & Davis, 1943). The frequency of impulses in a mammalian auditory nerve fibre signals the intensity of sound in a particular range of the auditory spectrum, each ending being coupled to a resonator which is sharply tuned (Gold & Pumphrey, 1948). The ear is thus incapable of registering the frequency of a sound pulse which contains less than a certain number of cycles of vibration.

An air vibration with the waveform of cicada song may produce several sensations to the human listener. Two frequencies are involved; the basic sound frequency (n), which in the species observed in Ceylon lies between 4000 and 7000 cyc./sec., and the pulse repetition frequency (m), which is between 120 and 600 cyc./sec. If the damping of the sound is high (i.e. if the number of cycles of vibration of the tymbal at each click is small) then the human ear registers mainly the pulse repetition frequency. If the damping is low, the most obvious frequency in the heard sound may be that of the tymbal vibration. This is particularly the case if the vibration does not die away completely before the next pulse and if the phase of the sound is preserved from pulse to pulse; such pulses are said to be ‘coherent’ (Text-fig. 15). If the pulses are coherent then the whole waveform may be considered as a single continuous sound vibration, pulse modulated at pulse repetition frequency. A Fourier analysis of such a waveform (which is in effect done by the human ear) resolves it into a number of superimposed sinusoidal oscillations; in addition to frequency n there are also present frequencies n + m, n—m, n + zm, n—zm, etc., at intensities depending on the exact waveform of the modulation. The human ear registers a musical chord, the quality depending on the harmonics.

Text-fig. 15.

Definition of ‘coherence’. Diagrammatic representation of two coherent pulses; the dotted section shows the preservation of phase between the two pulses.

Text-fig. 15.

Definition of ‘coherence’. Diagrammatic representation of two coherent pulses; the dotted section shows the preservation of phase between the two pulses.

If the pulses are not coherent, but nevertheless contain a number of cycles of oscillation, the human ear registers a ‘noise’, with components at a large number of frequencies including both the sound and modulation frequencies. The sensation is likely to be a sibilant hiss with the much lower modulation frequency superimposed as a faint tone. If there are few cycles of oscillation in each pulse, the main sound heard is the modulation frequency with harmonics of this note.

This digression is necessary in order that the observations reported in the next section may be fully appreciated.

The patterns of song

Magnetic tape recordings of cicadas singing in the field were made for nine species in Ceylon. The distribution and habits of these species are described more fully elsewhere (Pringle, 1954b); here we are mainly concerned with the physiological mechanism by which the patterns are generated.

Pl. 11 shows low-speed oscillograms of the song patterns, made from magnetic tape recordings. The significant parameters of the song are summarized in Table 2.

TABLE 2.

Analysis of the free song of nine species of cicada from Ceylon

Analysis of the free song of nine species of cicada from Ceylon
Analysis of the free song of nine species of cicada from Ceylon

Myers (1929) makes the point that the songs of cicadas are ‘no mere mechanical products ground out to a pattern pre-determined by the structure of the soundorgans in the species concerned. The sound-organs themselves are superficially as much alike as a number of violins collected at random ; the songs are as diverse as the tunes which may be produced from the instruments by different players. ‘This view is entirely confirmed by the present study. Small differences in the morphology of the nine species were found, but are not in general sufficient to warrant detailed description. The basic machinery of sound-production consists always of three functional elements:

  • (a) The tymbals and tymbal muscles, directly responsible for the generation of the pulses of sound.

  • (b) The tensor muscles, controlling the pulse repetition frequency and the volume of sound.

  • (c) The muscle complex controlling the resonance of the air sacs.

The varied patterns of song are produced by different patterns of excitation of the nerves supplying these three muscle groups.

The species may now be discussed individually.

Pur ana campanula

This small, slender species (body length 21 mm., width at front of abdomen 6 mm.) is considered first, since its song is the most distinctive to the human ear. The rhythmic unit lasts about 2-5 sec. and comprises a phrase with faffing amplitude and pulse repetition frequency (Pl. 11, Table 2). The song has a peculiar bell-like quality, the human ear registering not the pulse repetition frequency but the sound frequency modulated at pulse frequency. High-speed oscillographic analysis of the sound (Text-fig. 16A, B) reveals that the damping is low and that the pulses are coherent. The sound heard is therefore a chord, containing frequencies n (4000 cyc./sec.), n + m (4280 cyc./sec.), n—m (3720 cyc./sec.) and harmonics. The listener naturally pays most attention to the top note of a chord and the sensation is of a note falling in pitch; the simultaneous fall in intensity completes the resemblance to a bell.

A recently killed specimen of this species was used for experiments on the resonance of the air sacs. The cavity could in this case be excited by plucking the chitin of the abdominal segments with the point of a needle, the sound being recorded on an oscilloscope with triggered time-base. Measurements gave the frequency as 4000 cyc./sec. ; the air sacs are therefore closely resonant to the frequency of tymbal vibration. This and the length of the cavity relative to its breadth probably account for the low damping of the pulses.

The coherence of phase between pulses requires explanation, since, according to the explanation of the myogenic rhythm given by Pringle (1954 a), the IN click of the tymbal occurs at the moment when the tension in the apodeme of the tymbal muscle reaches the critical tension for the click as activation redevelops after the quick-release of the previous cycle ; the phase of the free oscillation of the tymbal has no obvious connexion with this cycle. The only satisfactory explanation of coherence is that the oscillation of the tymbal induces a small oscillatory tension in the apodeme which sums with the tension developed by the reactivation of the muscular contractile mechanism. In this way the instant of the IN click will be related to the phase of the previous oscillation, and coherence between pulses is preserved.

Some degree of coherence between pulses probably occurs in some other cicadas in which the damping is sufficiently low for the free oscillation of the tymbal not to have died away completely before the next IN click. But in no other species observed in Ceylon is it sufficiently marked to produce the auditory sensations given by the song of P. campanula. It may be noted here that the song of a Japanese species, Tanna japonensis, structurally related to Pur ana, has been described (Hearn, 1900, quoted by Myers, 1929) as having a bell-like quality.

Platypleura octoguttata

This species provides an extreme contrast to Purana campanula in the degree of damping and coherence of the sound pulses (Text-fig. 16, C-H). In Platypleura octoguttata the sound is quenched almost completely between pulses and the most noticeable sound to the human ear is that of the pulse repetition frequency. The song comprises a five-note sequence (Pl. 11) which can best be rendered in musical notation (Table 2). At the fourth note there is a rise in pitch of about a semitone with a drop in amplitude, and sometimes a further increase in the damping. The rhythm of the song is remarkably regular, varying only in the length of the final note (5, Table 2) which may be more prolonged than is shown in the figures.

Text-fig. 16.

Coherence of pulses in the free song. A and B, Purana campanula, coherent pulses; C and D, Platypleura octoguttata, pulses not coherent; E to H, P. octoguttata (another insect); E and G, note 5 of the song (Table 2); F and H, note 4 of the song. I, calibration for A to F, i msec. ; J, calibration for G and H, 1 msec.

Text-fig. 16.

Coherence of pulses in the free song. A and B, Purana campanula, coherent pulses; C and D, Platypleura octoguttata, pulses not coherent; E to H, P. octoguttata (another insect); E and G, note 5 of the song (Table 2); F and H, note 4 of the song. I, calibration for A to F, i msec. ; J, calibration for G and H, 1 msec.

High-speed oscillograms of the song of this species (Text-fig. 16C-H) provide clear evidence that both the IN and OUT clicks of the tymbal are effective during full song; the pulses are associated in pairs and the waveform resembles that obtained during laboratory experients in which tymbal clicks were induced by electrical stimulation. The frequency of contraction of the tymbal muscle is therefore half the pulse frequency, namely 440-480/sec.

There is no intermission of activity of the tymbal muscle during the song, the apparent breaks in the record of Pl. 11 being due merely to a great reduction in amplitude ; the rhythm of pulses is preserved. The abdomen is compressed during the fourth note in which the pulse repetition frequency is raised. Probably the tensor muscle is partially relaxed at this time and the air sacs compressed, reducing the critical tension necessary to click the tymbal and bringing the cavity out of resonance. The overall drop in amplitude of sound bears this out, as does the relative reduction in the amplitude of sound at the IN click (Text-fig. 16 E-H’). It will be recalled that experiments on the related P. capitata (Text-fig. ra) showed that contraction of the tensor muscle increased the energy dissipated at the IN click by increasing the curvature of the tymbal. At the apparent breaks in the song the tensor muscle may be completely relaxed for short periods, so that the clicks are very much less powerful.

P. octoguttata was the only one of the nine species which was observed to sing with a different rhythmic pattern when the male was in close proximity to a female A small number of these insects were confined on a bush in a garden in Colombo, and on several occasions a male produced what can only be called the ‘courtship’ song (Text-fig. 17). The volume of sound is considerably less than in the full song and the rhythm is different, consisting of a regular slow alternation of two notes each lasting about 2-5 sec., one being louder than the other. The courtship song is terminated by a loud shriek, upon which the male attempts to mount the female which may or may not remain still and allow copulation to take place. The pulst repetition frequency does not change appreciably during the phrases of the court ship song, but rises during the whole performance. It was measured in on< specimen as 560/sec. (double pulses) and is therefore higher than in the full song as would be expected if the tensor muscle is relaxed. The loud and soft alternations are accompanied by expansion and contraction of the abdomen, presumably bringing the air sacs into and out of resonance.

Text-fig. 17.

Platypleura octoguttata. The end of the courtship song; the two portions of the record are contiguous. Time marker 0-5 sec.

Text-fig. 17.

Platypleura octoguttata. The end of the courtship song; the two portions of the record are contiguous. Time marker 0-5 sec.

The fundamental sound frequency is higher in P. octoguttata than in any other species recorded, in spite of the fact that the insects are large (body length 25 mm., width at front of abdomen 12 mm.). In P. capitata, which is little larger (body length 29 mm., width at front of abdomen 12 mm.), the sound frequency is 4400-4700 cyc./sec. The two different specimens of P. octoguttata used for the records of Text-fig. 16 gave values of 5700 and 7100 cyc./sec.; other specimens gave 5400 and 5650 cyc./sec. It is possible that the insect giving 7100 cyc./sec. was abnormal in the sclerotization of its tymbal. Resonance of the air sacs was not measured in this species.

Platypleura westwoodi

This is a slightly smaller species than P. capitata or P. octoguttata, having a body length of 23 mm. and a width at the front of the abdomen of 10 mm. The song was heard on a number of occasions but was recorded only once (Pl. 11, Table 2). Again it seems that the tymbal muscle is continuously active ; the pattern is a rapid amplitude modulation at 17/sec. Visual observation during song was not possible, but presumably either the abdominal or the tensor muscles are responsible for the modulation.

Terpnosia ridens

Terpnosia species have a much narrower abdomen than Platypleura-, Terpnosia ridens measures 27 mm. in body length, but the greatest width of the abdomen is only 8-5 mm. Examination of living specimens of this and other species of the genus suggests that there is less capacity for abdominal dilatation than in Platypleura and that the abdomen is almost fully extended throughout life. The song pattern of Terpnosia ridens consists of a complex rhythm of pulse groups (Pl. 11, Table 2), with no measurable change in pulse repetition frequency. This suggests that the pattern is one of impulses in the tymbal nerves and that the insect is using the ‘on-off switch’ to its muscular rhythmic mechanism, rather than, as in Platypleura, leaving the switch on and modulating the rhythmic frequency by means of accessory muscles.

The short notes of the pattern (semiquavers, Table 2) contain from two to ten clearly double pulses and may represent single nerve impulses; the long note, in which the repetition frequency is 265/sec. (muscle contraction frequency), has a rising amplitude at the start and a suggestion that the IN click is more powerful than the OUT, so that there may here be a simultaneous contraction of the tensor muscle. Faster oscillograms of different portions of the song are shown in Text-fig. 18.

Text-fig. 18.

Terpnotia ridens. Four sections of the song pattern (see Pl. 11 and Table 2). Time marker 50 cyc./sec. A, second short group of semiquavers; B, middle of the long group of semiquavers; C, long note; D, quavers.

Text-fig. 18.

Terpnotia ridens. Four sections of the song pattern (see Pl. 11 and Table 2). Time marker 50 cyc./sec. A, second short group of semiquavers; B, middle of the long group of semiquavers; C, long note; D, quavers.

The basic rhythm of Terpnosia ridens contains 47 ± 2 notes of different duration, and is constant throughout the species; no variation outside this range was heard on three occasions on which populations of the insect were observed. The song is produced by each individual for a period of about a minute and is taken up by others, so that a wave of sound passes over a considerable area of forest. If a singing insect is disturbed it flies only a short distance and the song may continue without interruption of the rhythm.

T. ridens, which is a new species, is described in full by Pringle (1954). IT differs very little in morphological characters from T. stipata, which is more widely distributed in Ceylon and whose song is described next. The morphological differences between these two species, so far as can be determined from the limited material so far available, are so slight that their separation would hardly be justifiable were it not for the differences in distribution and above all in the pattern of song. Since the song is a necessary part of the mating behaviour, the appearance of a new inherited pattern could readily lead to the establishment of a separate interbreeding group, with morphological differences appearing only later.

Terpnosia stipata

Body length 35 mm.; width at front of abdomen 11 mm.

Again in this species there is evidence of intermittent excitation of the tymbal muscle. The song (Pl. 11, Table 2) consists of the repetition of a phrase lasting 0 4 sec. in which the amplitude and pulse repetition frequency increase. In each period of song the gaps between the phrases become less and less distinct, so that finally the insect is singing continuously with only slight amplitude modulation but with the cycle of pulse repetition frequency still present. Abdominal movement accompanies the song pattern, but is not very marked. The volume of sound is high.

Measurements of the resonant frequency of a dead specimen by flicking the abdominal segments with a needle gave the value of 4400 cyc./sec. The waveform of the sound produced by this species is often far from sinusoidal (Text-fig. 19). In sections where it is sinusoidal the frequency is about 5200 cyc./sec., but elsewhere in the record the fundamental is at about 2250 cyc./sec. with a large component of second harmonic at 4500 cyc./sec. Evidently the tymbal is capable of vibrating in several modes. The strengthening ribs are more heavily sclerotized than in most other species examined, and it is possible that its main mode of vibration is at about 2250 cyc./sec., the second harmonic being boosted by the resonance of the cavity. To the human ear the song has a peculiarly strident quality, which might be expected if the note at 2250 cyc./sec. is accompanied by its octave and other overtones. A similar impression is given by the song of T. ridens.

Text-fig. 19.

Terpnosia stipata. Fast oscillograms showing the sound waveform with high harmonic content. Time marker 1 msec.

Text-fig. 19.

Terpnosia stipata. Fast oscillograms showing the sound waveform with high harmonic content. Time marker 1 msec.

The pulse repetition frequency starts at about 120/sec. and rises to 290/sec. in each phrase. Later, when the emission of sound has become practically continuous, the pulses become double, giving a frequency of nearly 600/sec., which is audible to the ear as a high whine accompanying the high frequency noise. This must indicate that both the IN and OUT clicks are contributing to the sound.

Rihana mixta

A large cicada: body length 37 mm., greatest width of abdomen 13 mm.

The morphology of the tensor muscle is different to that of the other species examined; its fibres end basally not directly oil the knob of the metathorax, but on a circular attachment disk resembling that of the tymbal muscle, and a short apodeme attaches this to the metathoracic knob. There is no reason to think that this alters the function of the muscle.

The song (Pl. 11, Table 2) consists of two or more short notes, followed by a period of continuous activity, resembling ‘hip-hip-hooray ‘cheering. There is always a period of silence between the notes; this and the quality of the sound, which is not strident, distinguishes it from Terpnosia stipata\ the duration of the short (‘hip’) notes is also longer (0-75 sec.). Although to the ear the short notes of the song resemble those of T. stipata in that the amplitude and apparent frequency of the pulses increases, high-speed analysis of the sound track reveals a difference. The ‘pulses ‘in this part of the song of Rihana mixta are, in fact, groups of pulses (Fig. 20-4), which merge into continuous activity at the end of the note and during the period of continuous song (‘hooray’). Each short (‘hip’) note must therefore be interpreted as a burst of impulses of increasing frequency in the tymbal nerve fibres, starting at about 30/sec. at which frequency the tymbal muscle gives merely one or a short sequence of contractions at each impulse, and rising to a frequency at which the muscle is maintained continuously active for a short time. There is an accurately co-ordinated elevation and expansion of the abdomen during the short notes—a good example of the way in which the tensor muscle is used to balance the raised mean tension in the tymbal muscle when this is strongly excited.

Text-fig. 20.

Rihana mixta. High-speed oscillograms of the free song. Time marker 50 cyc./sec. A, consecutive record of a single ‘hip’ note, showing the composite nature of the ‘pulses’. The dots which have been added over the time marker indicate the instants of occurrence of impulses in the tymbal nerves, according to the interpretation given in the text. B and C, two positions of the ‘hooray’ note, showing occasional frequency-doubling during periods when both the IN and OUT clicks of the tymbal emit sound pulses.

Text-fig. 20.

Rihana mixta. High-speed oscillograms of the free song. Time marker 50 cyc./sec. A, consecutive record of a single ‘hip’ note, showing the composite nature of the ‘pulses’. The dots which have been added over the time marker indicate the instants of occurrence of impulses in the tymbal nerves, according to the interpretation given in the text. B and C, two positions of the ‘hooray’ note, showing occasional frequency-doubling during periods when both the IN and OUT clicks of the tymbal emit sound pulses.

The lack of synchronism between impulses in the tymbal nerves on opposite sides, which was found in a fatigued preparation and is discussed on p. 536, is not detectable in normal song; a single rhythm of pulses always occurs, the maximum frequency of contraction of the tymbal muscle being about 295/sec. The apparent irregularity of Text-fig. 20 B, C arises from the fact that both the IN and OUT clicks of the tymbal emit sound pulses of equal amplitude for part of the time and for the rest one of them is much reduced; the reason for this is not clear, but it is a constant feature of the song of this species.

No measurements were made of the resonance of the air sacs.

Cryptotympana exalbida

Another large cicada; body length 37 mm., greatest width of abdomen 14 mm. The song (Pl. 11, Table 2) has a regular rise and fall in amplitude and pulse repetition frequency every 0-5 sec., accompanied, at the instants of greatest amplitude, by a number of short breaks in the continuity of sound emission. This produces a noise somewhat resembling a knife-grinder. The probable explanation in terms of the neuromuscular mechanism is that the accessory muscles are excited in a 0-5 sec. rythm, and the tymbal muscle excitation (frequency of impulses in the tymbal nerve fibres) reduced momentarily to the point where individual nerve impulses evoke a short train of muscular contractions. If this explanation is correct the situation is the inverse of that found in Rihana mixta, in which the maxima of excitation of the tymbal and accessory muscles occur at the same instant in the pattern of song ; here the minimum of tymbal muscle excitation coincides with the maximum of tensor muscle excitation. The only tape recording which was made is not very suitable for high-speed oscillographic analysis and the estimate of muscle contraction frequency given in Table 2 must be accepted with caution.

Discussion of the song patterns

The nine species recorded in Ceylon reveal a great diversity of song types, and the analysis suggests that the pattern of muscular co-ordination is equally varied. The literature on cicada song from other countries indicates that a similar variety is found elsewhere, but the Ceylon species seem to cover most of the types of song found in the family with the exception of the ‘wing-clacking’, which Myers (1929) noted in Melampsalta cingulata. This is an entirely different method of producing sound which is found only in some cicadas and is probably a true stridulation produced by the rubbing of the wings together or on the sides of the body ; a comb-like striated mesonotal ridge which is plucked by a wing-fold has been described in the subfamily Tettigadinae and may represent a development of this accessory method of sound production. Unfortunately the descriptions given by earlier workers without the advantages of oscillographic recording are not sufficiently precise to allow the patterns of song to be analysed in the way attempted in this paper, and no systematic classification of the songs of different subfamilies or genera is possible from a survey of the literature. If recordings become available from a wider range of species, such a classification will be interesting and may help with the taxonomy of the family.

The accurate muscular co-ordination involved in the song patterns indicates a degree of inherited nervous organization which is, at first sight, remarkable, but may in fact be no more complex than is needed to produce particular types of locomotor activity. What is unusual in the animal kingdom is that such a diversity of patterns of co-ordination should have evolved within one family with so little modification of the basic anatomical plan. Fossil cicadas closely resembling present-day forms have been found in Tertiary rocks and doubtful specimens from the Cretaceous (Myers, 1929 review); the family is thus of similar antiquity to the birds, animals which may be compared with cicadas in the way in which complex behaviour patterns have evolved without much change in the structure of the vocal organs.

The tympanum of the auditory organ has already been mentioned in the list of structures. It is a delicate membrane situated on each side ventro-laterally under the operculum, and is composed of two layers, the external cuticle and the wall of the air sac, which are sufficiently closely apposed to form interference fringes. An apodeme connects the tympanum to the auditory capsule (Text-fig. 4) containing a large number of chordotonal sensilla, and from these the auditory nerve runs through the base of the chitinous V forward to the ganglion.

Vogel (1923), who made a detailed study of the anatomy and histology of the organ, describes and figures in the male of Huechys incamata A. & S. a tensor tympani muscle running from the dorsal rim of the tympanum to the tergum on each side. This muscle is not present in the mature adult male of Platypleura capitata, in which also the dorsal rim of the tympanum is formed by the lateral arms of the chitinous V. The muscle is well developed in the female (Pringle, 1955) and was also found in a recently emerged male in which the chitin was incompletely hardened. It seems likely that it is reabsorbed in the mature male of this species after serving to stretch the chitin of the tympanal rim; in P. capitata the lateral arms of the chitinous V become firmly attached to the abdominal terga as the chitin hardens (Text-fig. 4), making any tensor muscle unnecessary. Possibly there are specific differences in this point of anatomy.

Vogel misses the function of the detensor tympani muscle, which he calls simply ‘lângsmuskel’, and he does not call attention to the weak point in the tympanal rim near the mid-ventral line (Text-figs. 2, 4) at which movement takes place when the muscle contracts, creasing the tympanum. Otherwise his results were confirmed in the present investigation.

Auditory nerve

Impulses were led off from the auditory nerve near to its entry into the ganglion. The cicada ear is extremely sensitive to high-pitched sounds, and since no soundproof room was available it is doubtful if true silence was ever obtained. There was always a background discharge in the nerve, and the impression was gained that any high-pitched sound which was audible to the human ear produced an increase in the nervous discharge.

The discharge obtained in the auditory nerve of a male when the sound of a cicada of the same species was played back through the loudspeaker of the magnetic tape recorder is shown in Text-fig. 21. Each pulse of sound produces a volley of impulses in the auditory nerve, usually with some after-discharge, and the pulses are followed in the nerve up to the highest pulse frequency used (93/sec.). Similar results were also obtained with females. The cicada ear evidently functions, like other insect auditory organs (Pumphrey, 1940), as a rectifier of the received sound, the impulse pattern corresponding to the modulation envelope. The sound frequency in cicada song is therefore that of a carrier wave, and the significant information for the insect is transmitted in the pulse pattern. No attempt was made to determine experimentally whether the auditory receptor of the male is tuned to the sound frequency characteristic of the species, a feature which would increase its sensitivity and selectivity; the fact that the tympana are coupled into the same resonant cavity as the tymbals might increase their response in this frequency range. In the female, where the air sacs are much smaller, this effect could hardly be present unless the tympana themselves contributed to the resonance.

Text-fig. 21.

Platypleura capitata. Electrical record (A) from the auditory nerve of a male when the sound emitted by another cicada (B) is played back through a loudspeaker. Time marker 50 cyc./sec.

Text-fig. 21.

Platypleura capitata. Electrical record (A) from the auditory nerve of a male when the sound emitted by another cicada (B) is played back through a loudspeaker. Time marker 50 cyc./sec.

Nerve to the tensor muscles

This nerve, which contains the motor fibres to the tensor and dorsal muscles, has also a larger number of small sensory fibres. In preparations made by bisecting the insect vertically with a razor cut passing just to one side of the mid-line a considerable resting discharge can be recorded from the cut distal end of the nerve. The endings appear to lie in the region of the skeletal knob on the metathorax to which is attached the tensor muscle. This area is sensitive to vibration, which increases the nervous discharge, but attempts to localize the endings more precisely by mechanical stimulation with the point of a needle were unsuccessful. It is possible that there are near here some mechanical sense organs, of much lower sensitivity than the tympanal chordotonal organs, which act as a monitoring system for the vibrations produced by the insect when it is singing.

Tymbal nerve

There are no sensory fibres in the tymbal nerves.

Discussion of sensory physiology and behaviour

The results reported in this section are of a preliminary nature only. There is no doubt that both sexes have a well-developed auditory sense; the fact that both males and females of a species are often found assembled into a small group in three or four trees shows that the sense is effective in bringing the sexes together. To what extent the insects react only to the song pattern of their own species is not known, but field observations indicate that the grouping is usually specific. It is noteworthy that where several species of cicada were found in the same locality there was always a clear difference in song pattern between them. P. capitata and Terpnosia ransonetti, neither of which has any pattern to the song, do not overlap in geographical range. Perhaps the most similar song patterns of overlapping species are Platypleura octoguttata and P. westwoodi’, everywhere where these two species were found together the individuals of the latter were closely grouped in a pure population.

Cicada song

Although the song is almost certainly the stimulus which leads to the assembly of both sexes, it is not established that it provides the females with the only means of locating the males, once they are in the right area. Both chemical and visual stimuli may be involved at the stage of close approach. That females do find the males was made probable by the observation (often made before) that a singing male is frequently accompanied on the branch by one or more females. On two occasions with different species a female was found unexpectedly in the net when a male was captured. It is also clear that the presence of a female does not always lead to a change in the behaviour of the male; the latter may continue to sing at full volume for a long time after females have arrived. Possibly the distinctive courtship song is not present in all species.

The final initiative in copulation comes from the male, at any rate in P. octoguttata, in which attempted mounting was observed. For this the stimulus is probably visual except perhaps in those species of cicada in which the female also emits sound by the ‘wing-clacking’ mechanism.

A point which was not settled satisfactorily is whether the singing male has any facility for monitoring his own song. Although the tympana are creased by the action of the detensor tympani muscles the very considerable vibration in the whole of the skeleton may still be adequate to excite the auditory endings. These or the sensory endings supplied by the tensor nerves could provide a reflex means of adjusting the pattern of co-ordination on the ‘reafferenzprincip ‘(v. Holst & Mittelstaedt, 1950). But it is difficult to see how any of these sense organs could give an indication to the insect of the amplitude of sound energy produced as opposed to the amplitude of vibration at pulse repetition frequency, unless there is such a wide range of thresholds in the auditory chordotonal sensilla that some of them are still within their working range at the intensities of sound produced by the insect itself. It may be that there is a purely instinctive pattern of muscular co-ordination which achieves tuning of the cavity without sensory feed-back ; further work is needed to settle this question.

The characteristics of the auditory organ—in particular its reception of merely the modulation envelope of the received sound—imply a lack of sensitivity to the relative phasing of the sound pulses. Coherence or its absence are thus features of the sound which carry no information to another insect. It is also doubtful if the degree of damping of the sound pulses will have much influence on the impulse pattern in the auditory nerve. The inevitable conclusion, that those features of cicada song which have most appeal to the naturalist are those least likely to be important in determining behaviour, may perhaps serve to encourage future workers on these insects to make good use of instrumental aids.

I am grateful to Prof. Koch and the staff and assistants in the Department of Physiology, University of Ceylon, for the help which they gave me during my stay. The visit was made possible by the award of a Leverhulme Research Fellowship and by grants from the H. E. Durham Fund of King’s College, Cambridge, and from the British Council. Part of the apparatus used was purchased with the aid of the Government Grants Committee of the Royal Society. Mrs B. Petrie was responsible for the drawings of cicadas in Text-fig. 1. Dr W. E. China has given me much help in the identification of species.

The magnetic tape recording of cicada songs was played to a meeting of the Society for Experimental Biology in London in January 1954. Gramophone records made from the tape have been deposited with the Colombo Museum. The tape may be borrowed on application to the author at the Department of Zoology, Cambridge.

Berlese
,
A.
(
1909
).
Gli Insetti
,.
Milan
.
Binet
,
A.
(
1894
).
Contributions a l’étude du système nerveux sous-intestinal des insectes
.
J. Anat
.,
Paris
,
30
,
449
.
Bullock
,
T. H.
(
1953
).
Properties of some natural and quasi-artificial synapses in polychaetes
.
J. Comp. Neurol
.
98
,
37
.
Carlet
,
G.
(
1877
).
Mémoire sur l’appareil musical de la cigale
.
Amt. Sei. nat. (zool.)
,
5
,
1
.
Cassertus
,
J.
(
1600
).
De vocis auditusque organis historia anatómica
.
Ferrara
.
Distant
,
W. L.
(
1906
).
The Fauna of British India. Rhynchota
, vol.
111
.
London
.
Felici
,
G.
(
1724
).
Lettera intomo al canto delle cicale
.
Giom. de’ Letterati d’Italia
,
36
,
56
.
Galambos
,
R.
&
Davis
,
H. J.
(
1943
).
The response of single auditory-nerve fibres to acoustic stimulation
.
J. Neurophysiol
.
6
,
39
.
Gold
,
T.
&
Pumphrey
,
R. J.
(
1948
).
Hearing. I. The cochlea as a frequency analyser
.
Proc. Roy. Soc. B
,
135
,
462
.
Haupt
,
H.
(
1929
).
Die Mechanik des Zikadenflugels und ihre Bedeutung fur den Flug
.
Z. wist. InsektBiol
.
24
,
73
.
Hearn
,
L.
(
1900
).
Shadowings
.
Boston
.
Hickernell
,
L. M.
(
1923
).
The digestive system of the periodical cicada, Tibicen septemdecim Linn. III. Morphology of the system in the nymph
.
Biol. Bull., Woods Hole
,
45
,
213
.
Hingston
,
R. W. G.
(
1922
).
A Naturalist in Himalaya
.
Boston
.
v. Holst
,
E.
&
Mittelstaedt
,
H.
(
1950
).
Das Reafferenzprincip ; (Wechselwirkungen zwischen Zentralnervensystem und Peripherie)
.
Naturwissenschaften
,
37
,
464
.
Landois
,
H.
(
1867
).
Die Ton-und Stimmapparate der Insekten in anatomisch-physiologisches und akustischer Beziehung
.
Z. wiss. Zool
.
17
,
105
.
Lloyd Morgan
,
C.
(
1886
).
Sound-organs of Platypleura capensis—Singetjie
.
Nature, Lond
.,
33
,
368
.
Lucas
,
A. H. S.
(
1887
).
On the sound-organs of the green cicada, Cyclochila Australasiae Donovan
.
Trans. Roy. Soc. Viet
.
23
,
173
.
Mayer
,
P.
(
1877
).
Der Tonapparat der Cicaden
.
Z. wiss. Zool
.
28
,
79
.
Myers
,
J. G.
(
1928
).
The morphology of the Cicadidae
.
Proc. Zool. Soc. Lond
. p.
365
.
Myers
,
J. G.
(
1929
).
Insect Singers
.
London
:
Routledge
.
Nastuk
,
W. L.
&
Hodgkin
,
A. L.
(
1950
).
The electrical activity of single muscle fibres
.
J. Cell. Comp. Physiol
.
35
,
39
.
Ossiannilsson
,
F.
(
1949
).
Insect drummers
.
Opuse, ent. (suppl
.),
10
.
Pierce
,
G. W.
(
1948
).
The Songs of Insects
.
Harvard University Press
.
Pringle
,
J. W. S.
(
1948
).
The gyroscopic mechanism of the halteres of Diptera
.
Phil. Trans. B
.
233
.
347
.
Pringle
,
J. W. S.
(
1949
).
The excitation and contraction of the flight muscles of insects
.
J. Physiol
.
108
,
226
.
Pringle
,
J. W. S.
(
1953
).
Physiology of song in cicadas
.
Nature, Lond
.,
172
,
248
.
Pringle
,
J. W. S.
(
1954a
).
The mechanism of the myogenic rhythm of certain insect striated muscles
.
J. Physiol
.
124
,
269
.
Pringle
,
J. W. S.
(
1954b
).
The songs and habits of Ceylon cicadas, with a description of two new species
.
Spolia zeylan
. (In the Press.)
Pringle
,
J. W. S.
(
1955
).
The structure and evolution of the organs of sound-production in cicadas. (In preparation
.)
Pumphrey
,
R. J.
(
1940
).
Hearing in insects
.
Biol. Rev
.
15
,
107
.
Reaumur
,
R. A. F. De
(
1740
).
Mémoires pour servir à l’Histoire des Insectes
,
5
,
145
206
.
Paris
.
Roeder
,
K. D.
(
1951
).
Movements of the thorax and potential changes in the thoracic muscles of insects during flight
.
Biol. Bull., Woods Hole
,
100
,
95
.
Snodgrass
,
R. E.
(
1921
).
The seventeen-year locust
.
Rept. Smithson. Inst
.,
1919
, p.
381
.
Vogel
,
R.
(
1923
).
Über ein tympanales Sinnesorgan, das mutmassliche Hôrorgan der Singzikaden. Z. ges. Anat. 1
.
Z. Anat. EntwGesch
.
67
,
190
.
Weber
,
H.
(
1928
).
Skelett, Muskulatur und Darm der schwarzen Blattlaus, Aphis fabae Scop
.
Zoológica
,
28
,
Heft
76
.
Weber
,
H.
(
1930
).
Biologie der Hemipteren
.
Berlin
:
Julius Springer
.

Oscillograms from magnetic tape recordings of Ceylon cicadas. Time marker (for all records) 0-5 sec.

Oscillograms from magnetic tape recordings of Ceylon cicadas. Time marker (for all records) 0-5 sec.